Atlas

Compare Models

Claude Sonnet 5.5 vs Muse Ember

Claude Sonnet 5.5 leads by 4.4 capability points (≈95% interval +1.9 to +6.9): a clear lead over Muse Ember · P(higher) ≈ 100%

Wins 47 of 50 shared benchmarks (3 lost)

Capability vs. relative cost
Claude Sonnet 5.5Muse EmberClosed weightsOpen weights

Benchmark wins47 – 3 · 50 shared

Comparing matching setups only (3 benchmarks excluded) · suites counted once
← Claude Sonnet 5.5Muse Ember →
Data: Artificial Analysis, Surge AI, Arena and 11 othersatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Ember minus Claude Sonnet 5.5.

Terminal-Bench-Science 0.1Science › Multidisciplinary53.320.0−33.3
LMArena Agent Arena Net Improvement (Steering Burden)Tool & computer use › General agents15.61.0−28.5
Humanity's Last Exam DiamondKnowledge & reasoning › General knowledge53.325.1−28.2
DAXBench v2Software › Data & ML86.262.8−23.4
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & medicine41.718.3−23.3
RiemannBench (November 2026 rescoring)Math76.853.6−23.2
Terminal-Bench 4.0 (Artificial Analysis)Software › Software engineering63.640.9−22.7
LMArena Code Arena WebDevSoftware › Frontend & design17731608−21.9
Artificial Analysis Intelligence Index v5.0.1Composite indexes54.536.0−18.5
DeepFinanceProfessional › Finance, accounting & tax54.636.5−18.1
HANDBOOK.md Agents v1.1Tool & computer use › Customer service & operations47.729.8−17.9
DAYJOB: Healthcare (October 2026 rescoring)Health & medicine34.017.2−16.8
Artificial Analysis Engineering IndexScience › Engineering58.541.8−16.7
LMArena Agent Arena Praise vs ComplaintTool & computer use › General agents30.7-2.7−16.7
APEX-1 (Mercor, mean score)Professional › General knowledge work53.970.4+16.5
DAYJOB: Finance (October 2026 rescoring)Professional › Finance, accounting & tax40.524.2−16.3
AA-Briefcase v1.1 EloProfessional › General knowledge work18151520−16.2
TaxCalcBench TY25Professional › Finance, accounting & tax18.02.0−16.0
HLAB-Vals-1.1Professional › Legal10.425.0+14.6
FrontierCode 2.0 (Artificial Analysis)Software › Software engineering41.727.2−14.5
Artificial Analysis Legal IndexProfessional › Legal52.238.5−13.7
GDP.xlsxProfessional › General knowledge work29.115.4−13.7
APEX-SWESoftware › Software engineering66.452.9−13.5
CritPt v2 (agentic, avg@5)Science › Physical sciences75.762.3−13.4
Artificial Analysis Economics IndexComposite indexes54.141.0−13.1
Artificial Analysis Healthcare & Medical IndexHealth & medicine54.241.3−12.9
SWE Atlas v1.1 - OverallSoftware › Software engineering68.356.0−12.3
Ramp SWE-Bench v2Software › Software engineering63.251.8−11.4
AA-Omniscience v1.1 IndexKnowledge & reasoning › General knowledge30.57.7−11.4
BuseyBench SVGPerception & media › Generation6.55.4−11.3
Artificial Analysis Finance & Accounting IndexProfessional › Finance, accounting & tax55.844.6−11.2
GBENCH One-shot Coding GScoreGames66.456.4−10.0
LMArena Agent Arena Confirmed SuccessTool & computer use › General agents17.6-2.0−9.8
Artificial Analysis Strategy & Ops IndexProfessional › General knowledge work58.148.7−9.4
EnterpriseBench: CoreCraftTool & computer use › Customer service & operations70.862.1−8.7
Vibe Code Bench v1.1Software › Frontend & design92.483.7−8.7
ComplexConstraints (November 2026 rescoring)Language & writing › Instruction following67.258.8−8.4
KradleBenchGames33.825.7−8.1
APEX-Agents 1.1Professional › General knowledge work83.775.7−8.0
Artificial Analysis Progress Index v1Composite indexes351329−7.6
FORTRESS Average Risk Score (ARS)Safety & alignment18.211.9+6.3
LMArena Agent Arena Steering BurdenTool & computer use › General agents14.84.8−5.0
Excel Modeling Benchmark v1.1Professional › Finance, accounting & tax75.972.2−3.7
AutomationBench-AA 1.0.6Tool & computer use › Customer service & operations71.868.1−3.7
SimpleBenchKnowledge & reasoning › General knowledge75.972.7−3.2
GDP.pdf (Artificial Analysis, All-pass)Professional › General knowledge work25.822.8−3.0
LMArena Agent Arena Bash ReliabilityTool & computer use › General agents11.96.6−2.6
GDP.pdf (November 2026 rescoring)Professional › General knowledge work51.249.2−2.0
τ³-Banking (Mercor run, bm25_grep retrieval)Tool & computer use › Customer service & operations54.053.3−0.7
LMArena Agent Arena Tool HallucinationTool & computer use › Tool calling1.00.90.0
Data: Artificial Analysis, Surge AI, Arena and 11 othersatlas.village.dev
Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…